Simple statistical approach presenting Partial Discharge data in large pool of MV motors in continuous and remote PD monitoring system

This paper is based on two decades of experience with continuous monitoring of Partial Discharge (PD) in medium voltage equipment including Rotating Machines. The technology uses multiple sensors distributed along monitored objects and 1 - 20 MHz frequency range. It allows for additional noise reduction and for localizing of PD activity sources. The importance of continuous PD monitoring vs. periodic PD measurements is discussed considering that PD activity can be very unstable and highly dependent on ambient conditions. Negative periodic PD tests in equipment can result in a false sense of security. Case study from field experience is presented to support the discussions. Advances in communication and computing have allowed Remote Monitoring and Continuous Monitoring to become very effective combinations to make sense of Partial Discharge diagnostics relieving owners of the burden of data analysis for intricate PD diagnostics. Alarms can be detected and analyzed in near real-time, and previously overlooked pre-failure conditions can be detected to avoid a critical failure. Remote data collection is safe and secure, experts are readily available, and portals and web pages provide easy access to accumulated data. Amount accumulating data allowed for statistical analysis and different presentation of the data to customers. Statistical approach gives users wide picture of where their equipment stands compare to a large pool of similar machines. Despite of evident flaws and introduced arbitrary rules statistical approach got acceptance and became standard tool for automatically generated reports accessible through secure WEB site to Remote Monitoring users.